See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/378022038

Introduction of Artificial Intelligence Literacy and Data Literacy in Computer

Science Teacher Education

Conference Paper · February 2024

DOI: 10.1145/3631802.3631851

CITATIONS READS

7 278

9 authors, including:

Viktoriya Olari Martin Kandlhofer

Freie Universität Berlin Austrian Computer Society
11 PUBLICATIONS 143 CITATIONS 27 PUBLICATIONS 948 CITATIONS

SEE PROFILE SEE PROFILE

Ulrike Lucke
Universität Potsdam
219 PUBLICATIONS 917 CITATIONS

SEE PROFILE

All content following this page was uploaded by Viktoriya Olari on 21 February 2024.

The user has requested enhancement of the downloaded file.

 (Authors' Draft)

Introduction of Artificial Intelligence Literacy and Data Literacy

in Computer Science Teacher Education
Viktoriya Olari Martin Reger Martin Kandlhofer
Till Zoppke Evgenia Samoilova [email protected]
Ralf Romeike Ulrike Lucke Österreichische Computer
[email protected] [email protected] Gesellschaft (OCG)
[email protected] [email protected] Vienna, Austria
[email protected] [email protected]
Freie Universität Berlin Universität Potsdam
Berlin, Germany Potsdam-Golm, Germany

Anna Sarah Lieckfeld Valentina Dagienė

[email protected] [email protected]
German Informatics Society Vilnius University
Berlin, Germany Vilnius, Lithuania

ABSTRACT 1 INTRODUCTION
The integration of artificial intelligence (AI) literacy and data liter- In order to thrive in a data-driven world, school students need to
acy into school education is increasingly proposed in educational become familiar with artificial intelligence (AI) from an early age
frameworks, leading to an urgent need to train computer science [10] and develop AI literacy and data literacy (AI&DL). The former
teachers in AI and data literacy. However, there is a lack of pro- is a set of competencies that enables individuals to understand and
fessional development programs for these skills. To fill this gap, critically evaluate AI technologies; communicate and collaborate
we developed a training concept and evaluated it with computer effectively with AI; and use AI as a tool online, at home, and in
science teachers from Germany, Austria, and Lithuania in three the workplace [5, 6]. The latter is the ability to collect, manage,
sessions. In the following, we present selected evaluation results. evaluate, and apply data, in a critical manner [8]. However, the key
to successful integration of AI&DL into school education is effective
CCS CONCEPTS teacher preparation. Despite the growing body of research on AI
education, there is a lack of knowledge about AI teacher training
• Computing methodologies → Artificial intelligence; • Applied programs for K-12 education, according to a recent literature review
computing → Education; • General and reference → Empirical [9].
studies. Through our research, we are taking a step toward research-
based teacher education in AI&DL. Informed by the requirements
KEYWORDS for professional teacher development programs that we gathered in
data literacy, artificial intelligence literacy, teacher education, action a dialogue with European stakeholders and by reviewing European
research, mixed methods educational policies, we developed a one-day professional develop-
ment program for in-service computer science (CS) teachers with
ACM Reference Format: no to some prior knowledge in AI&DL and evaluated it with CS
Viktoriya Olari, Till Zoppke, Ralf Romeike, Martin Reger, Evgenia Samoilova, teachers from Germany, Austria, and Lithuania under the guidance
Ulrike Lucke, Martin Kandlhofer, Anna Sarah Lieckfeld, and Valentina of the following two research question: (1) What is the effect of
Dagienė. 2023. Introduction of Artificial Intelligence Literacy and Data the designed program on teachers’ perceived competence to incor-
Literacy in Computer Science Teacher Education. In 23rd Koli Calling In- porate AI&DL into their teaching and on their understanding of
ternational Conference on Computing Education Research (Koli Calling ’23), AI&DL concepts?; (2) To what extent are teachers able and willing
November 13–18, 2023, Koli, Finland. ACM, New York, NY, USA, 2 pages. to incorporate AI&DL content introduced in the program into their
https://doi.org/10.1145/3631802.3631851 teaching, and what are the potential barriers?

2 DEVELOPMENT OF THE CS TEACHER

Permission to make digital or hard copies of part or all of this work for personal or TRAINING
classroom use is granted without fee provided that copies are not made or distributed We developed a one-day (7-hour) teacher training concept that
for profit or commercial advantage and that copies bear this notice and the full citation
on the first page. Copyrights for third-party components of this work must be honored. includes two components: (1) content knowledge (based on the
For all other uses, contact the owner/author(s). Dagstuhl triangle framework [2]) and (2) pedagogical knowledge. In
Koli Calling ’23, November 13–18, 2023, Koli, Finland terms of content knowledge, we focused on fundamental paradigms
© 2023 Copyright held by the owner/author(s).
ACM ISBN 979-8-4007-1653-9/23/11. of rule-based AI, supervised, unsupervised and reinforcement learn-
https://doi.org/10.1145/3631802.3631851 ing (AI-related content) and the data lifecycle (DL-related content),
 Koli Calling ’23, November 13–18, 2023, Koli, Finland Olari et al.

as these are recurring themes in international AI&DL frameworks Teacher training for Data Literacy & Computer Science compe-
[5, 6]. For the pedagogical knowledge, we based the training on the tences".
didactic biplane which is commonly used for CS teacher training
[11]. We used research-based open-source unplugged learning ma- REFERENCES
terials for machine learning and rule-based AI published under a [1] Jordan Allison. 2023. Classifying the Characteristics of Effective Continuing
Professional Development (CPD) for Computer Science Teachers in the 16-18
free license by Wissensfabrik [7]. For the data lifecycle, we chose Sector. ACM Transactions on Computing Education 23, 2 (Jun 2023), 1–30. https:
the computer-based activity in Orange3 [4], also published by the //doi.org/10.1145/3582275
Wissensfabrik [7]. [2] Torsten Brinda and Ira Diethelm. 2017. Education in the Digital Networked World.
In Tomorrow’s Learning: Involving Everyone. Learning with and about Technologies
and Computing (IFIP Advances in Information and Communication Technology),
Arthur Tatnall and Mary Webb (Eds.). Springer International Publishing, Cham,
3 RESULTS AND CONCLUSION 653–657. https://doi.org/10.1007/978-3-319-74310-3_66
[3] C. Hong and Salika Lawrence. 2011. Action Research in Teacher Education:
We evaluated the impact of the training concept in three sessions Classroom Inquiry, Reflection, and Data-Driven Decision Making. Journal
with 70 CS teachers from Germany, Austria and Lithuania using of Inquiry and Action in Education (2011). https://www.semanticscholar.org/
paper/Action-Research-in-Teacher-Education%3A-Classroom-and-Hong-
a pre- and post-evaluation survey, an AI&DL self-assessment and Lawrence/29433e4d13847d2d2a94a13a517c18104e88ca94
knowledge test, and semi-structured personal interviews. The re- [4] Janez Demšar, Tomaž Curk, Aleš Erjavec, Črt Gorup, Tomaž Hočevar, Mitar
sults of the evaluation showed that the training in all three countries Milutinovič, Martin Možina, Matija Polajnar, Marko Toplak, Anže Starič, Miha
Štajdohar, Lan Umek, Lan Žagar, Jure Žbontar, Marinka Žitnik, and Blaž Zupan.
had some positive impact on teachers’ perceived competences on 2013. Orange: Data Mining Toolbox in Python. Journal of Machine Learning
how to use AI content in class, as well as teachers’ understanding of Research 14 (2013), 2349–2353. http://jmlr.org/papers/v14/demsar13a.html
AI concepts introduced in the training. Results were mixed for DL [5] Duri Long and Brian Magerko. 2020. What is AI Literacy? Competencies and
Design Considerations. Proceedings of the 2020 CHI Conference on Human Factors
content: improvements in this area showed high variability with in Computing Systems, 1–16. https://doi.org/10.1145/3313831.3376727 [Online;
no apparent increase after training. Participants in all three coun- accessed 2023-01-31].
[6] Tilman Michaeli, Ralf Romeike, and Stefan Seegerer. 2023. What Students Can
tries reported that the training served as a valuable introduction to Learn About Artificial Intelligence - Recommendations for K-12 Computing
AI and DL. However, the depth and complexity of certain topics, Education. In Towards a Collaborative Society Through Creative Learning, Therese
especially in the realm of AI, were challenging for participants. Keane, Cathy Lewin, Torsten Brinda, and Rosa Bottino (Eds.). Springer Nature
Switzerland, Cham, 196–208. https://doi.org/10.1007/978-3-031-43393-1_19
Exercises using the Orange3 tool consistently received mixed feed- [7] Tilman Michaeli, Stefan Seegerer, and Ralf Romeike. 2022. Modul KI-B3: Schlag
back. While some teachers found it suitable for teaching, others den Roboter und Modul KI-B4: Von Daten und Bäumen, Reihe IT2School -
found it too complex and criticized its graphical interface and lack Gemeinsam IT entdecken. https://www.wissensfabrik.de/it2school/. Accessed:
2023-06-23.
of programming features as a deterrent for more advanced students. [8] Chantel Ridsdale, James Rothwell, Mike Smit, Michael Bliemel, Dean Irvine, Dan
In terms of teachers’ perceived ability and willingness to integrate Kelley, Stan Matwin, Brad Wuetherick, and Hossam Ali-Hassan. 2015. Strategies
and Best Practices for Data Literacy Education Knowledge Synthesis Report.
the AI&DL into their teaching, participants highlighted several dif- https://doi.org/10.13140/RG.2.1.1922.5044
ficulties. While teachers felt more prepared to integrate the AI&DL [9] Ismaila Temitayo Sanusi, Solomon Sunday Oyelere, Henriikka Vartiainen, Jarkko
content presented in the training into their classrooms after at- Suhonen, and Markku Tukiainen. 2023. A systematic review of teaching and learn-
ing machine learning in K-12 education. Education and Information Technologies
tending the training, integration would require further engagement 28, 5 (May 2023), 5967–5997. https://doi.org/10.1007/s10639-022-11416-7
with the material. A single 7-hour course was not sufficient to en- [10] UNESCO. 2021. Beijing Consensus on Artificial Intelligence and Education. (Jul
sure the integration of DL and AI into the classroom, especially if 2021). https://unesdoc.unesco.org/ark:/48223/pf0000368303
[11] Diethelm Wahl. 2013. Lernumgebungen erfolgreich gestalten: vom trägen Wissen
teachers had not taught these subjects prior to the training. Teach- zum kompetenten Handeln (Successfully designing learning environments: from
ers expressed a strong desire for further training, concrete course inert knowledge to competent action) (3. auflage mit methodensammlung ed.).
Verlag Julius Klinkhardt, Bad Heilbrunn.
plans and materials tailored to specific grade levels to streamline
integration into teaching and reduce their preparation workload,
which corresponds with prior research on professional training
programs for CS teachers [1]. The main barriers for CS teachers
to implement AI&DL in their classrooms were lack of knowledge
and time constraints such as current workload. Through our find-
ings, we hope to raise awareness of teacher training within the AI
education research community and shed light on how to design
and evaluate the impact of the professional development program
for AI&DL. As we are following the action research approach [3]
in developing, implementing, and evaluating the teacher training
concept, we would like to discuss our interim results within the
research community. This exchange could be particularly valuable
as we also explore ways to effectively evaluate the long-term impact
of the training in future work.

ACKNOWLEDGMENTS
We would like to thank Erasmus+ Programme of the European
Union for financial support as a part of the project "TrainDL -

View publication stats